Michael Stonebraker

Michael Stonebraker

ID:40085143

大小:622.63 KB

页数:37页

时间:2019-07-20

Michael Stonebraker_第1页
Michael Stonebraker_第2页
Michael Stonebraker_第3页
Michael Stonebraker_第4页
Michael Stonebraker_第5页
资源描述:

《Michael Stonebraker》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、WhatDoes‘BigData’MeanandWhoWillWin?MichaelStonebrakerTheMeaningofBigData-3V’s•BigVolume—Businessstuffwithsimple(SQL)analytics—Businessstuffwithcomplex(non-SQL)analytics•BigVelocity—Drinkfromafirehose•BigVariety—Largenumberofdiversedatasourcestointegrate2BigVolume-LittleAnalytics•Well

2、addressedbydatawarehousecrowd•WhoareprettygoodatSQLanalyticson—Hundredsofnodes—Petabytesofdata•Buttherehasbeenadramaticseachangeinthelastdecade!3TraditionalRowStoreArchitecture(circa2005)Record1Record2Record3Record4E.g.DB2,Oracle,Sybase,SQLServer,Greenplum,Netezza,Teradata,…ColumnSto

3、reArchitecture(now)CCCOOOLLLUUUMMMNNN12KE.g.Vertica,SybaseIQ,MicrosoftPDW,DB2-BLU,RedShift,…TheFacts•Acolumnstoreis50Xfasterthanarowstore!!!!•Thesuccessfulvendorswillallrun“sharenothing”multi-nodecolumnstoresforscalability•Mosteverybody–exceptOracle–hasgottenthereoverthelastdecadeOra

4、cleExadataisnot:acolumnstoreascalableshared-nothingarchitecture6BigData-BigAnalyticsTheNextBigThing!•Complexmathoperations(machinelearning,clustering,trenddetection,….)—Theworldofthe“quants”andthe“rocketscientists”—Mostlyspecifiedaslinearalgebraonarraydata•Adozenorsocommon‘innerloops

5、’—Matrixmultiply—QRdecomposition—SVDdecomposition—Linearregression7BigData-BigAnalyticsAnExample•Considerclosingpriceonalltradingdaysforthelast5yearsfortwostocksAandB•Whatisthecovariancebetweenthetwotime-series?(1/N)*sum(Ai-mean(A))*(Bi-mean(B))8NowMakeItInteresting…•Dothisforallpair

6、sof4000stocks—Thedataisthefollowing4000x1000matrixStockttttttt….t12345671000S1S2…S40009ArrayAnswer•Ignoringthe(1/N)andsubtractingoffthemeans….Stock*StockT•TrythisinSQLwithsomerelationalsimulationofthestockarray!!!!10DataScientistJobDescription•Cleanandcuratethedata(majorityofhistimer

7、ightnow–forwardpointertoVariety)•Until(tired){Datamanagementoperation(s);Complexanalyticsoperations(s);}11UseCaseRequirements•Complexanalytics—Covarianceisjustthestart—Definedonarrays—Graphsarejustsparsearrays!!!•Datamanagement—Leaveoutoutliers—Justonsecuritieswithamarketcapover$10B•

8、Scalabilityt

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。